openvinotoolkit / openvinotoolkit/nncf

[Good First Issue][NNCF]: Support not transposed weight for data-aware weight compression methods

Open
#3,494 8 comments 0 reactions 1 assignee View on GitHub

@yuanjua is already working on this.

Since May 15, 2025.

good first issue
Dominant language
Python
Stars
1.2k
Forks
305
Avg merge
1d 9h
Merged PRs (30d)
27

Description

Context

Matmul operation in OpenVINO assumes an implicit shape alignment for input arguments. It applies transpositions specified by optional transpose_a and transpose_b attributes: OV spec.
Currently, weight compression in NNCF does not support transpose_b=False.
Here's the test.
Potentially, it affects Mixed-Precision, AWQ, Scale Estimation, GPTQ and Lora Correction algorithms.

What needs to be done?

The task is to enable data-aware weight compression methods (Mixed-Precision, AWQ, Scale Estimation, Lora Correction, GPTQ) for models with matrix multiplications having not transposed weight.

test_compression_with_transpose shouldn't raise an error for transpose_b=False

Example Pull Requests

https://github.com/openvinotoolkit/nncf/issues/3230
https://github.com/openvinotoolkit/nncf/pull/3296

Resources
Contact points

@ljaljushkin

Ticket

No response

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.